A Materials Science Perspective of Midstream Challenges in the Utilization of Heavy Crude Oil.

An increasing global population and a sharply upward trajectory of per capita energy consumption continue to drive the demand for fossil fuels, which remain integral to energy grids and the global transportation infrastructure. The oil and gas industry is increasingly reliant on unconventional deposits such as heavy crude oil and bitumen for reasons of accessibility, scale, and geopolitics. Unconventional deposits such as the Canadian Oil Sands in Northern Alberta contain more than one-third of the world's viscous oil reserves and are vital linchpins to meet the energy needs of rapidly industrializing populations. Heavy oil is typically recovered from subsurface deposits using thermal recovery approaches such as steam-assisted gravity drainage (SAGD). In this perspective article, we discuss several aspects of materials science challenges in the utilization of heavy crude oil with an emphasis on the needs of the Canadian Oil Sands. In particular, we discuss surface modification and materials' design approaches essential to operations under extreme environments of high temperatures and pressures and the presence of corrosive species. The demanding conditions for materials and surfaces are directly traceable to the high viscosity, low surface tension, and substantial sulfur content of heavy crude oil, which necessitates extensive energy-intensive thermal processes, warrants dilution/emulsification to ease the flow of rheologically challenging fluids, and engenders the need to protect corrodible components. Geopolitical reasons have further led to a considerable geographic separation between extraction sites and advanced refineries capable of processing heavy oils to a diverse slate of products, thus necessitating a massive midstream infrastructure for transportation of these rheologically challenging fluids. Innovations in fluid handling, bitumen processing, and midstream transportation are critical to the economic viability of heavy oil. Here, we discuss foundational principles, recent technological advancements, and unmet needs emphasizing candidate solutions for thermal insulation, membrane-assisted separations, corrosion protection, and midstream bitumen transportation. This perspective seeks to highlight illustrative materials' technology developments spanning the range from nanocomposite coatings and cement sheaths for thermal insulation to the utilization of orthogonal wettability to engender separation of water-oil emulsions stabilized by endogenous surfactants extracted during SAGD, size-exclusion membranes for fractionation of bitumen, omniphobic coatings for drag reduction in pipelines and to ease oil handling in containers, solid prills obtained from partial bitumen solidification to enable solid-state transport with reduced risk of damage from spills, and nanocomposite coatings incorporating multiple modes of corrosion inhibition. Future outlooks for onsite partial upgradation are also described, which could potentially bypass the use of refineries for some fractions, enable access to a broader cross-section of refineries, and enable a new distributed chemical manufacturing paradigm.

Introduction

Fossil fuels continue to play a central role in meeting global energy needs, from powering the electric grid to heating habitats and fueling the transportation infrastructure that underpins our unprecedented age of global connectivity.[1] Oil refineries process crude oil to a diverse slate of products, not just transportation and heating fuels, but also chemical feedstock and bituminous components of asphalt infrastructure, reflecting a complex entanglement of the oil and gas industry with chemicals and road infrastructure. The oil and gas industry (and by proxy, the global economy) is increasingly reliant on unconventional deposits such as heavy crude oil and bitumen for reasons of accessibility, scale, and geopolitics.[2,3] Unconventional deposits in Canada and Venezuela contain more than one-third of the world’s viscous oil reserves and have emerged as vital linchpins to meet the energy needs of growing and rapidly industrializing populations.[4] Crude oil, sweet and sour (classified based on sulfur content), light or heavy (depending on molecular weight and classified based on specific gravity; heavy oil typically has an API gravity < 22.3°, specific gravity > 920 kg/m3), is transported to refineries using deepwater ports, rail links, and pipelines. In turn, an extensive network of conduits, the very arteries of global economies, carries processed fuel, chemical feedstock, plastics, and fertilizer from refineries and associated chemical plants to manufacturing and population centers across the world. The high viscosity of heavy crude oil and bitumen presents a substantial impediment to their extraction, midstream transportation, and processing. As such, challenges associated with the vexing rheological properties of these fuels have far-ranging implications for energy security, economic resilience, and manufacturing. As an illustration of the magnitude of this problem, the Canadian Oil Sands produce 4 million barrels a day, primarily sour heavy crude, amounting to 22% of total US imports.[5] The contentious debate regarding the construction of new pipelines in North America underscores the burning need for fundamental scientific innovations as well as their effective translation to viable midstream technologies.[6,7] In this perspective article, we discuss the challenges and opportunities for materials science in midstream processing, storage, and transportation of heavy crude oils, emphasizing potential solutions for more effective utilization of hydrocarbon resources in the Athabasca region of North America.

Oil Recovery in the Canadian Oil Sands

Since the global spike in oil prices in the 1970s, significant attention has focused on the recovery of heavy crude oil from unconventional deposits. As a result of advancements in horizontal drilling, enhanced oil recovery, and associated advancements in process intensification, the production of unconventional deposits has become economically viable in large measure. The Canadian Oil Sands are a prominent example. Production is expected to increase from the current 4 million barrels a day by up to an additional 1.5 million barrels per day over the next decade—almost all of the additional production bound for consumption in the United States. Tertiary enhanced oil recovery (EOR) methods have played a central role in the emergence of viable production of bitumen and are broadly classified into two categories, thermal and nonthermal.[2,3,8] Nonthermal methods include cold production, chemical flooding, and miscible displacement.[2] In chemical flooding, the surface energy of the formations and deposits are modified through the addition of an amphiphile, which thereby promotes the mobilization of oil. In miscible displacement, gases are injected into the cores and the density of oil is modified to ensure mobilization. These methods are plagued by the high costs of chemical additives and the extended periods of time required to initiate the production of heavy oil. In the Canadian Oil Sands, nonthermal methods have had limited success (recovery rates limited to 12–15%) and are not as widely used as thermal alternatives.[3] The most efficient EOR methods induce flow through thermal mechanisms that fundamentally modify the viscosity of heavy oil.[2] One such method is cyclic steam stimulation (CSS), wherein a singular well is injected with high-pressure steam at 300–340 °C, allowed to soak over a few days to weeks, and then the oil is pumped to the surface. As the production well flow begins to ebb, the process is repeated, typically for at least 15 cycles.[9] A modification of this approach that has been widely adopted and represents the state-of-the-art is steam-assisted gravity drainage (SAGD), which yields a recovery factor of approximately 60%.[9] In this process, as sketched in Figure , two stacked horizontal wells are drilled into the formation. The topmost injection well is used to administer high-pressure steam at temperatures in the range of 90–250 °C, promoting emulsification, facilitated by endogenous surfactants, which modifies the viscosity of heavy oil and enables its mobilization.[10] The secondary, production well, drilled still deeper into the reservoir, then collects the drained complex emulsions from the steam chamber; the recovered emulsions span the range from oil-in-water, water-in-oil, oil-in-water-in-oil, and water-in-oil-in-water multinary mixtures. Once pumped to the surface, complex emulsions need to be separated. The produced water is deoiled, desalinated, and recycled back to the SAGD process. The separated heavy oil is transported (sometimes with partial upgradation such as hydrogen treatment) to refineries, which requires resolution of all of the challenges incumbent from its high viscosity and sulfur content.

Figure 1

Schematic depiction of the opportunities for materials science solutions throughout the SAGD process for enhanced oil recovery of viscous oil in the Athabasca region of North America. The large temperature variations, high pressures, and harsh corrosive environments present a complex set of challenges for materials’ design.

Materials’ Chemistry Challenges in the SAGD Process

Figure sketches the steps involved, highlighting the opportunities for materials science innovations in the design of membranes, surface coatings, cement sheaths, pipelines, and solutions for solid-state transport. High fluid temperatures, frigid external temperatures, high subsurface pressures, flammable liquids, and harsh corrosive environments require the design of resilient, functional materials that are tailored to resist degradation in extreme “far-from-equilibrium” environments.[11−13]

Despite rapid technological advancements that have rendered SAGD methods an economically viable approach for oil production in the Canadian Oil Sands, the production of oil by such methods remains considerably more expensive as compared to extraction of sweet light crude in the Permian Basin or from plays in the Persian Gulf. As such, there is considerable interest in improved recovery rates through the incorporation of chemical additives, energy-efficient thermal cementing to prevent heat loss during SAGD (or incorporation of thermal insulation more generally across the SAGD infrastructure), more optimal separations of the extracted emulsions, protection of base metal components exposed to harsh corrosive environments, and the design of solutions to midstream transport that mitigates myriad current challenges with pipeline and railcar transportation.[14−19]

Heavy oil and bitumen are shipped using heated tankers, rail cars, trucks, and pipelines, oftentimes requiring dilution with light hydrocarbons to obtain fluids with rheological properties amenable to transportation. As such, the transportation of bitumen entails a massive energy expenditure and requires the installation and maintenance of a substantial thermal infrastructure. The need for diluents further adds substantial cost to midstream transport; ca. 30% of pipeline capacity in North America is tied up in the unproductive flow of diluents, light hydrocarbons, added simply to modify the rheology of bitumen. Figure A shows that whereas synthetic and conventional oil exports have remained fairly constant over the last 20 years, there has been a massive increase in Canadian exports of diluted bitumen to the United States from 170 000 barrels per day in 2000 to 2.4 million barrels per day in 2019.[20] As a result of the high viscosity and low surface tension of bitumen, the maintenance and cleaning of transportation vessels incur a considerable cost, results in substantial unrecovered residues, and engenders safety hazards for those engaged in maintaining transportation equipment.

Figure 2

(A) Comparison of pipeline tolls and railcar transportation costs from Alberta to different refinery operations in North America depicting the daily refinery capacity (inset), reproduced with permission from Oil Sands Magazine https://www.oilsandsmagazine.com/market-insights/crude-oil-pricing-differentials-why-alberta-crude-sells-at-deep-discount-to-wti#references.[21] (B) Yearly export volumes based on oil type from conventional, synthetic, and dilute bitumen from 1990 to 2019, information on the license can be found at https://open.canada.ca/en/open-government-licence-canada.(20)

Notwithstanding uncertainty introduced from COVID-19, the US Energy Information Administration pegs future US transportation energy needs at around 26 quadrillion BTUs in 2050 with over 80% projected to derive from hydrocarbon fuels.[1] As noted above, 22% of total US imports are derived from Canada (Figure A).[21] Pipelines are the most efficient mode to transport crude oil, yet all existing pipelines to the Unites States are at capacity. The US refineries closest to the Canadian Oil Sands in the Midwest (Petroleum Administration for Defense Districts (PADD)-2, PADD2) have a surfeit of Canadian Oil (100% of imports). A clear solution would be to transport heavy oil produced in excess of PADD2 capacities to PADD3 refineries in the Houston Gulf Coast, which have a total refining capacity of 10 million barrels a day (Figure B); these are some of the most complex refineries capable of handling heavy crude, offer the best prices for heavy and sour Canadian crude, and provide access to deepwater ports for transportation of a complex slate of products across the world. These refineries were historically constructed to handle heavy crude from Venezuela and Mexico; supplies from both countries have plummeted in recent years leading to attractive pricing for Canadian heavy and sour crude. In the absence of adequate pipeline capacity, railcar transportation has emerged as the primary means of transportation between Western Canada and the US Gulf Coast. Canada Energy Regulator and Statistics Canada estimate that 400 000 barrels of Western Canada Heavy Sour Crude from Alberta are transported daily by railcar to refineries in the United States; this number has risen sharply since railcar transport was nonexistent until 2012 and is expected to further increase with the cancellation of the Keystone XL pipeline. As shown in Figure , railcar transportation is considerably more expensive as compared to other shipping methods (adding $15–22 to the cost of a barrel of oil delivered to the refinery). For both pipeline and rail transport, there is thus considerable interest in the design of surface modification approaches that will enable less fouling of the infrastructure, reduce the amount of diluents required, and enable transportation at ambient temperatures. Alternative midstream strategies have focused on solid-state transport such as through encapsulation in polymer wastes or through reconstitution of bitumen to form solid prills wherein lighter fractions are coated by asphaltene shells.[22−24]

Regardless of the mode of transportation, the vast midstream infrastructure is primarily constructed from structural steel and thus prone to corrosion. Corrosion has an impact of $2.5 trillion globally and specifically in the United States; the annual cost induced by the railways amounts to $11.16 million.[25,26] Corrosion further risks spillage of oil in environments, which can have a devastating impact on vulnerable ecosystems.[27] A 2018 National Academies Report notes that pipeline infrastructure is more robust and has greater safeguards in place as compared to railcar transportation. Forthcoming sweeping sets of reforms will entirely alter the landscape of railcar transportation in North America by putting in place stringent specifications for rail cars as all rail cars must be built to meet DOT-117 requirements or retrofitted to comply with the transportation of Class 3 flammable liquids, ethanol, and unrefined petroleum products.[28] These new regulations address safety concerns such as top fittings and thermal insulation. Existing rail cars not meeting these standards are due to be phased out by 2029.[29]

In this perspective article, we outline the formidable and distinctive materials science challenges faced by the midstream oil industry in the handling of heavy oil, delineate the underlying scientific principles to possible solutions, and discuss technological advancements based on materials’ chemistry innovations that represent candidate solutions to these challenges. Such advances in materials chemistry are of pivotal importance to ensuring the economic viability of the Canadian Oil Sands and to mitigate deleterious impacts on the environment. In compiling this perspective article, we have focused on some recent contributions, which are discussed in the context of the broader literature. In the interest of presenting a succinct narrative, we have emphasized illustrative examples instead of attempting a comprehensive review. The subsequent sections discuss thermal insulation at the wellhead and in well cementing, membrane-assisted separations, corrosion protection, and solutions for midstream transportation of viscous oil. In addition to a discussion of the fundamental principles and promising outcomes in the field, we provide a prospectus for future opportunities.

Coatings and Cement Sheaths for Thermal Insulation

SAGD being a thermal recovery process is extremely energy intensive as it requires handling of hot fluids across different stages of operation,[30] e.g., ground transportation of steam from boilers to the wellhead, steam injection through the vertical well bore, maintaining the temperature of the steam chamber, and transportation of viscous heavy oil to refineries.[31,32] All of these steps are critical and direct determiners of the overall productivity and cost of production. Therefore, minimizing energy losses at every step is pivotal to the efficiency and cost-effectiveness of the SAGD process.[33,34] Operational modifications aimed at reducing energy losses include the injection of a solvent intermittently with steam (solvent-assisted SAGD),[35,36] injecting noncondensable gases with steam (NCG-SAGD),[37,38] coinjecting surfactant solutions, performing foam-assisted SAGD (FA-SAGD),[39,40] and drilling injection and production wells perpendicular to each other (cross SAGD).[41] Alternative strategies have focused on increasing energy efficiency through improved materials selection and design without altering the basic SAGD operational process. This entails reducing energy losses by insulating all surfaces from where the heat is lost by conduction, convection, and radiation.[42,43]

Reliance on heating in SAGD oil production indeed makes thermal energy the major contributor to production costs. Hence, any loss of thermal energy input during the operation of the SAGD process has a direct impact on the economic viability of the process. The energy losses occur primarily through the surfaces of pipelines as a result of the high thermal conductivity of pipeline materials (mostly structural steels) and large thermal gradients across these conduits given frigid external temperatures in the Athabasca region.[19] Thermal conductivity is a measure of heat transfer efficiency within a material under the influence of a temperature gradient. When a material is exposed to a higher temperature at one surface, the temperature gradient that develops between the hot and the relatively cold surfaces leads to heat energy transfer across the material as per Fourier’s law[44]where the heat flow, Q, is expressed as a function of the surface area, A; ΔT is the temperature difference between the hot and cold surfaces; d is the separation between the surfaces; and k is the thermal conductivity of the material. In the case of SAGD pipes that transport hot fluids, the high thermal conductivity of steel (50 W/(m·K)) and large temperature gradients (10–40 °C external temperatures and ca. 180–250 °C fluid temperatures) result in a large propensity for heat loss. A decrease of fluid temperature is accompanied by a sharp increase of viscosity of heavy oils,[45] which can give rise to severe transport limitations, thereby increasing the demand for steam and resulting in higher steam-to-oil ratios (SOR: barrels of steam required to produce a barrel of oil),[46] and eventually reducing the efficacy of the SAGD process. Therefore, it is imperative to reduce the thermal conductivity of these surfaces to render the process more energy efficient. As shown in Figure A, blanketing hot surfaces with an insulation coating reduces the overall thermal conductivity of the insulated surface as chemically dissimilar materials in contact at the heat exchange surface act as serially connected thermal resistors. The overall thermal conductivity (ktotal) can be expressed aswhere Ts is the internal temperature of the steel pipe, Tc is the external temperature of the insulating coating, and ds and dc are the thicknesses of the steel pipe and the insulating coating, respectively. Materials that impede ballistic heat transport by phonon scattering at many interfaces and are of low inherent thermal conductivity with enclosed void spaces show promise for the thermal insulation of pipelines.

Figure 3

(A) Schematic depiction of heat transfer across an insulated steel surface; (B) graphical representation of heat transfer from oil-transporting pipelines to the surroundings; (C) schematic representation of heat transport pathways within a composite structure comprising glass microspheres embedded within an epoxy matrix; and (D) cross-sectional scanning electron microscopy (SEM) image of a steel surface coated with an insulating polymer coating with embedded glass bubbles (GB) with chemical structures of candidate high-temperature polymers in the inset. Panel (B) Reprinted from Journal of Petroleum Science and Engineering, 160, Yuan, Q.; Wu, C.; Yu, B.; Han, D.; Zhang, X.; Cai, L.; Sun, D., Study on the Thermal Characteristics of Crude Oil Batch Pipelining with Differential Outlet Temperature and Inconstant Flow Rate, 519–530, 2018, with permission from Elsevier.[47] Panel (C) licensed with permission under CC BY 4.0.[48]

The means of applying insulation varies depending on the intended application of a pipeline. For example, steel pipelines used for ground transportation of crude oil need to be insulated to avoid flow constraints that would increase oil viscosity and require higher pumping pressures (Figure B).[47] Hence, insulation on such pipelines requires maintaining the inner temperature at 90–130 °C, sufficiently high to maintain a flow of crude oil (<500 cP).[31,47] In contrast, pipelines that transport steam, both above-ground (boilers to the wellhead) and underground (steam injecting wells), must maintain inner temperatures at much higher values, typically in the range of 180–290 °C. In general, above-ground pipe insulation methods are more versatile than underground methods owing to their relative ease of access. Common insulation methods include insulating blankets wrapped around the outer surface of pipelines; such blankets are commonly constituted from polyurethanes, polyethylene, and porous calcium silicates optionally combined with glass fibers, mineral wool, and alumina silicate fibers—all materials exhibiting low thermal conductivity and encapsulating substantial void space.[48−50]

Much recent effort has focused on embedding rigid hollow structures within polymeric media to achieve a combination of low density and high porosity, which impedes heat transfer and reduces thermal conductivity down to values <1 W/(m·K). As illustrated in Figure C, effective heat transfer within such composite structures is governed by three fundamental pathways: (1) gaseous convection and conduction within hollow insertions, (2) solid conduction across percolative networks, and (3) thermal radiation on the surface of hollow insertions.[48] Unfortunately, the glass-transition temperature of epoxy polymers is below 150 °C, which precludes their widespread application in pipelines wherein operating temperatures can be well in excess of this value. As such, considerable attention has focused on the design of high-temperature polymers, ceramic coatings, metal matrix composites, and ceramic composites.[51,52] We have recently developed polybenzimidazole- and polyimide-based composite coatings exhibiting low thermal conductivity (as low as 0.059 W/(m·K)) at high working temperatures (250–400 °C) that can be directly applied onto wellhead components and pipelines that transport steam.[53] The record low thermal conductivity values are obtained by embedding hollow glass bubbles inside the polymer matrix, which introduces void spaces and interfaces for phonon scattering. Figure D exemplifies a cross-sectional view of the hybrid coating.

An intriguing strategy for mitigating delamination of coatings from thermal stresses generated as a result of differentials in coefficients of thermal expansion (CTE) is based on the incorporation of negative thermal expansion (NTE) materials such as ZrV2O7 and HfV2O7 within high-temperature polymers. HfV2O7 exhibits isotropic NTE with a CTE = −6.7 × 10–6 °C–1 in the 130–700 °C range that is particularly relevant to SAGD operations. As an example of this approach, incorporation of NTE material HfV2O7 in a high-temperature polybenzimidazole resin enables as much as a 67.3% reduction in thermal stress at a relatively low loading of 27 vol %, suggesting a means of accessing zero thermal expansion nanocomposite coatings that can sustain large thermal variations encountered in SAGD processes.[53]

For the purpose of insulating well bores, commonly used methods include an annulus gas blanket or insulated tubing. An annulus gas blanket is formed by injecting gaseous nitrogen into the well bore.[49] Injected gas pushes the steam further into the well bore and creates a low thermal conductivity sheath that limits heat loss. Additionally, vacuum-insulated tubing (VIT) is gaining popularity as a means of reducing thermal losses at a SAGD well bore.[54,55] Adopting a concentric or eccentric tube-in-a-tube conformation (Figure A) yields a vacuum around the steam injection pipe, thereby substantially mitigating heat losses.[55] A concentric arrangement is preferred over an eccentric arrangement as the physical contact between the tubing and the casing in eccentric arrangement leads to some conductive heat losses. Even though VIT is estimated to increase the steam quality by 4.6% as compared to bare tubing,[54] it requires vacuum conditions as high as 10–5–10–6 Torr to effectively eliminate convective heat losses.[32]

Figure 4

(A) Cross-sectional diagram of concentric (up) and eccentric (down) tube arrangements in VIT; (B) schematic illustrating a modified cement formulation incorporating polymer-functionalized halloysite fillers; (C) transmission electron microscopy (TEM) image of halloysite nanotubes; (D) crystal structure of halloysite nanotubes; (E) thermal conductivity of unmodified and hydroxyethylcellulose halloysite nanotube (HEC-HNT)-modified cement at different weight loadings of HEC-HNT; and (F) compressive strength evolution of unmodified and HEC-HNT-modified cement upon prolonged thermal cycling (Note: WC = with CaCl2 accelerant, # = number of thermal cycles). Panel (B) licensed under creative commons CC BY.[18] Panel (D) reprinted from Materials and Design, 57, Albdiry, M. T.; Yousif, B. F., Role of silanized halloysite nanotubes on structural, mechanical properties and fracture toughness of thermoset nanocomposites, 279–288, 2014, with permission from Elsevier.[56] Panels (E, F) republished with permission of Engineering Research Express from An Evaluation of the Reduction of Heat Loss Enabled by Halloysite Modification of Oil well Cement, Udayakantha, M.; Cho, J.; Liu, K.-W.; Mukhopadhyay, A.; Gupta, S.; Hong, C. Y.; Banerjee, S., 1, 2, 2019.[19]

As an alternative approach, we have recently demonstrated that thermal energy losses in vertical injection wells can be substantially reduced by endowing thermal insulation capabilities to oil-well cement sheaths used to adhere steam pipes to the well bore. As shown in Figure B, we have embedded hydroxyethylcellulose-modified halloysite nanotubes (HEC-HNTs) within the cement matrix. HNTs are hollow nanotubes (Figure C) with a general formula of Al2Si2O5(OH)4 and enclose a lumen surrounded by aluminate interiors and silicate exteriors (Figure D). Transient hot bridge thermal conductivity measurements (Figure E) show the highest thermal resistance for a filler loading of 2 wt %. This formulation reduces the overall thermal conductivity of oil-well cement by as much as 76% without compromising mechanical properties.[18,19]Figure F demonstrates that the 28 days compressive strength of HEC-HNT-modified cement samples is well within the acceptable range for industrial operation (8–20 MPa) even after accelerated thermal cycling (heating at 250 °C for 20 h and cooling at 20 °C for 4 h) for 20 consecutive cycles. Indeed, the diminution in compressive strength with thermal cycling is lower for HEC-HNT-modified cement as compared to unmodified cement. As such, the introduction of HEC-HNTs introduces disparate interfaces and voids, resulting in increased phonon boundary scattering, thereby reducing thermal energy loss and ensuring steam quality while protecting the cement sheath from thermal fatigue.

Even though the primary role of these coatings is to provide thermal insulation, they are often called upon to provide additional functionality. For example, underground SAGD pipelines are prone to corrosion.[57,58] As such, there is a need to develop insulating coatings that can withstand high temperatures on the steel surface while being resistant to corrosion. Corrosion is particularly detrimental in VIT arrangements as it increases the tendency for hydrogen permeation, which at concentrations higher than the sorption capacity of embedded getters will destroy the vacuum.[32] In addition, the geographic locations of SAGD operations in the United States and Western Canada impose stringent constraints on the performance expectations of such insulating coatings. Specifically, coatings need to withstand substantial temperature gradients and continuous cyclic thermal stresses. Additional challenges arise from the need to paint or spray coatings onto existing operational pipelines with minimal surface preparation. As such, in addition to functional performance, a key criterion is adhesion to base metal surfaces. In summary, the design of insulating coatings and sheaths that are cost effective imbue high thermal resistance and corrosion inhibition and are readily applicable to operational pipelines remains an area of active research.

Separation of SAGD Emulsions

Complex oil–water emulsions are pumped to the surface in the SAGD process (Figure ) and comprise water-in-oil, oil-in-water, and concentric o–w–o and w–o–w droplet dispersions. The bitumen–water interface is stabilized by endogenous surfactants comprising resins, asphaltenes,[59,60] naphthenic acids (oligomeric carboxylate surfactants),[61] and minor quantities of linear alkanoic acids, tetra-acids, sulfonic acids, asphaltenic acids, etc., as well as mineral solids such as clays, calcite, silica, and quartz.[10,62−67] In this section, we provide a succinct discussion of the chemical origins of emulsion formation as well as the possible molecular mechanisms of destabilizing emulsions, describe the concept of differential wettability underpinning membrane separations, and review recent progress in affecting the separation of complex emulsions. Finally, we delineate strategies for the creation of tailored membranes for the effective separation of complex emulsions.

The mechanism of interface stabilization is depicted in Figure . Heavy fractions of bitumen contain polar functional groups and heteroatoms that interact via hydrogen bonding, intermolecular π–π interactions, covalent coordination bonds, Lewis acid–base interactions, electrostatic interactions, and van der Waals’ forces.[68] It is typical for resin fractions to accumulate at the oil–water interface as a result of faster adsorption kinetics owing to a significant reduction in interfacial tension.[69] Conversely, asphaltenes display slow but irreversible adsorption kinetics and aggregate within interfacial films through intermolecular π–π interactions, which prevent droplets from coalescing.[70] Mineral solids coupled with aggregated asphaltenes can create biwettable interfaces on particle surfaces, augmenting the stability of surfactant-stabilized emulsions by forming remarkably stable Pickering emulsions.[65,71] The cation exchange capacity of solid particles, porosity of minerals, crystallite and aggregate dimensions of inorganic solids, and molecular size of asphaltenes are important parameters governing asphaltene adsorption on solid particle surfaces.[72] The exceptional stability of such emulsions derives from a combination of small droplet dimensions, highly negative ζ-potential values,[63] and an abundance of amphiphilic species.[73] Under these conditions of emulsification, the produced heavy-oil–water emulsion is exceptionally stable. As such, stable SAGD emulsions must be actively disrupted to effectively separate aqueous and hydrocarbon fractions.[63,65,73]

Figure 5

(A) Schematic depiction of soil mineral and asphaltene–resin aggregate interaction in the creation of bitumen–water Pickering emulsions. (B) Wetting regimes of a droplet placed on a rough surface. (C) Methodology for membrane-assisted emulsion separation.

In general, thermodynamically unstable emulsions are kinetically stabilized by interfacial steric or electrostatic interactions. The Gibbs–Duhem equation[74] (eq ) describes three contributions to surface free energy: an entropy term, an interfacial energy term, and a composition term.At constant temperature and compositionwhere dGσ is the differential change in Gibbs free energy, Sσ is the entropy of the system, dT is the differential change in temperature, A is the interfacial area, n is the number of moles of component i with chemical potential μ, and γ is the interfacial tension. At constant temperature and composition, the entropy and compositional terms tend to zero (eq ) and interfacial tension is constant (eq ) and positive for surfaces; as such, if the interfacial area increases, the change in Gσ is positive; as such, extended interfaces between phases are thermodynamically disfavored.

In contrast, kinetic stabilization depends on the absorbed layer of surface-active agents on the O–W interface and the chemistry of the surface-active agent.[75] If the surface-active agent is ionic in nature, the stabilization occurs through electrostatic forces that provide the necessary repulsion between like charges on the double layers of two droplets.[76,77] Nonionic surfactants/amphiphiles stabilize the dispersed phase through steric hinderance. The concentration of long chains increases in the area where two droplets interact. Due to this increase in the local concentration of the surface-active agent in this area (overlapped region), the dispersing phase starts diffusing to the overlapped region which eventually causes the reduction of local concentration of long chains, thereby diminishing the overlapped region and finally separating droplets from each other. It is worth noting that despite their thermodynamic instability, kinetic stabilization enables SAGD emulsions to persist for a period of months and years. The purpose of demulsification processes is to disrupt interfacial interactions, culminating ultimately in phase separation across a hierarchy of length scales. At the nanoscale, smaller droplets merge to form larger droplets, contributing to droplet growth by Ostwald ripening.[78] Laplace pressure, Δp, within the droplet is a differential pressure term defined in eq , where r is the radius of the spherical droplet and γ is the interfacial tension. Laplace pressure governs the diffusion of surface species from the interface of smaller droplets (higher Laplace pressure) to larger droplets (lower Laplace pressure), resulting in greater stability of the merged droplet.On the scale of microns, when two or more droplets of the dispersed phase combine to form an aggregate, the result is flocculation.[79] The dispersed droplets can take on thermal energy from continual motion, buoyancy forces, gravitational forces, and applied mechanical forces. When the available energy surpasses a threshold value, the interface between droplets is eliminated, merging the flocculated droplets into a single larger droplet.[80] At still greater length scales, differentials in density drive the gravitational separation of droplets, which over a period of time leads to the formation of a distinct phase boundary. Sedimentation is observed when the dispersed phase is denser than the dispersing phase. In contrast, if the dispersed phase is lighter than the dispersing phase, the process is termed creaming. All of these processes need to be systematically controlled to engender macroscopic phase separation.

SAGD-extracted oil–water emulsions that are pumped to the surface are typically first sent to a water knock-out tank that uses density differentials to engender gravity-based separation of oil- and water-rich phases as well as mineral particulates.[81] Thermally driven cracking of oil-rich emulsions remains a ubiquitous means of separating the aqueous and oil fractions of SAGD emulsions, but alternatives are being actively explored because of their energy-intensive nature. The water-rich emulsion is delivered to a skim tank for primary treatment of wastewater, where the thin film of oil floating on the surface of water is removed. Secondary treatment involves the removal of suspended microscopic/nanoscopic oil droplets by induced gas flotation to facilitate coalescence and subsequently removal of the demulsified oil component using an oil removal filter. Finally, the tertiary treatment removes water hardness (calcium and magnesium) and dissolved silica by the addition of limewater at elevated temperatures known as warm lime softening. Here, limewater reacts with dissolved carbon dioxide forming carbonate and bicarbonate. This reaction shifts the equilibrium of carbonate species in the system, thereby precipitating large quantities of calcium carbonate when it exceeds its solubility product. Magnesium is removed in the same processes by a double displacement reaction. Lastly, ion-exchange absorption removes residual hardness in water by exchanging them with sodium ions. Finally, water is transported to the steam generator for reinjection.[82]

Produced water generated as a byproduct of the SAGD process has been a major liability since the advent of this technology. As an unconventional oil recovery method, SAGD has a produced-water-to-oil ratio (PWOR) as much as 4-fold that of conventional recovery methods.[83] However, the separated water is usually not suitable for reinjection, disposal (discharge limit <40 ppm of oil-in-water), or recycling and requires extensive further processing.[84,85] In particular, reuse of water from the SAGD process has stringent water quality constraints to protect boilers from scaling and corrosion.

Emerging membrane-based technologies are a promising alternative to the current flow trains used for the treatment of produced water or destabilization of water-in-oil emulsions. The interplay of the adhesive and cohesive forces on a multiphase interface can lead to preferential repulsion/attraction of a particular phase at a surface boundary.[86,87] Two broad regimes of wettability can be distinguished (Figure B).[88,89] The Wenzel regime[89] describes droplets that have found their way to a global free-energy minimum wherein the liquid resides on the surface and enters grooves, thereby reducing contact angles to <90°. Equation defines the contact angle in the Wenzel regime, where θw. γLV, γSL, and γSV are defined as the surface energy of liquid–vapor interface, solid–liquid interface, and solid–vapor interface, respectively. r is defined as the roughness factor (ratio of the actual contact area of solid to the projected area of contact on the solid–liquid interface).A droplet in the Cassie–Baxter regime[88] corresponds to a metastable configuration wherein the liquid is suspended over a composite air–solid interface with minimal or no permeation within the textural elements. A modified Cassie–Baxter[90] calculation of the observed contact angle (θCB) is expressed in eq , where rf is the roughness factor (ratio of the actual contact to the projected area of contact on the solid–liquid surface) and fSL is the fraction of the solid contact surface to the apparent total contact area of unit. The modulation of interfacial interactions is achieved by the surface modification of membranes to reduce/increase the surface free energy, enabling selective permeation of liquids along surface tension gradients.[91] Surface texturation plays an important role by amplifying the inherent propensity of a surface to wet or repel a liquid droplet.[91,92]Interfacial interactions can initiate and propagate the process of destabilizing emulsions through a sequence of processes: disruption of molecular interactions, flocculation to form droplets, mesoscale ripening and coalescence of droplets, stabilization of a phase boundary, and selective permeation of the continuous phase through the membrane.

Figure C depicts the aspects of the development of membrane-aided emulsion separation technologies. Membrane design must be tailored to the characteristics of the emulsion, such as the presence of suspended particulate matter and its size distribution, nature, and concentration of surfactants and the relative ratio of oil and water. Interfacial interactions can be tuned by modulating the pore size, surface texturation, and surface energy of the membrane, as well as by adding de-emulsifiers or combining with thermal processes. Process design, spanning the range from the selection of flow geometries and flow profiles to backwashing sequences and multiple looping, is furthermore of the utmost importance for commercial deployment of membrane technology, which is predicated on optimization of metrics such as separation efficiency, flux rate, resistance to fouling, and operational lifetime.

Surface engineering approaches for oil–water separation typically combine selective texturation of surfaces (through embedding inorganic nanoparticles, cellulosic fibers, or structured polymers) in conjunction with chemical modification of membrane surfaces. A review by Dickhout et al. outlined the fundamental principles underpinning separation mechanisms such as surface charge, surface roughness, concentration polarization, crossflow velocity, and transmembrane pressure.[93] Although an extensive literature has focused on the relative performance metrics of membranes for oil–water separations, fundamental materials’ design principles remain less explored.[94−98] Membrane function involves multiscale interactions between liquid flows and surfaces involving textural elements, surface functional groups, pore dimensions, adhesion, and adsorption-induced modification of interfaces.

In general, a membrane architecture comprises a porous support, which serves as the primary structural bulwark. The surface of the porous support is further imbued with multiscale texturation by embedding nanometer- or micron-sized morphological architectures. Surface energy is then modulated by the chemical functionalization of the textured surface. The structural support is oftentimes a periodic metal[99−101] or metal–alloy mesh,[17,102,103] ceramic membrane,[98,104,105] or woven microfiber textile (constructed from natural[106−108] or synthetic fibers).[109−111] Base support membranes define initial pore size and provide the primary means of size exclusion, which is further augmented through surface modification to enhance the separation of emulsions. Texturation enhances the intrinsic affinity of the surface without changing its free energy.

A variety of metal oxides that exhibit multiscale roughness are resistant to corrosion and are accessible at relatively low cost have been used as textural elements. Titanium oxides, aluminum oxide,[82,112−114] and silica[102,115,116] nanoparticles are the most widely used components of metal-oxide oil–water separation membranes. ZnO tetrapods have furthermore attracted recent interest because of their distinctive tetrapodal morphology, which provides a means of imbuing 3D mesoscale texturation.[15,17,117] Deposition of chemical/electrochemical reaction products of copper on the copper mesh has further been used to generate uniform micro/nanoscale roughness on porous surfaces.[99−101,118]

Rare-earth metal-oxide nanoparticles (CeO2–)[119] have been proposed to be intrinsically hydrophobic morphological elements for inducing surface roughness. It has been suggested that octet electrons in the outer orbitals (5s2 5p6) of rare-earth atoms completely shield 4f orbitals, resulting in drastically reduced polar interactions of water with the surface. As per this hypothesis, reduced cohesion of water droplets renders lanthanide oxide surfaces hydrophobic.[120] This mechanism has been challenged in more recent work wherein X-ray photoelectron spectroscopy and dynamic water contact angle measurements indicate that rare-earth oxides are intrinsically hydrophilic but rendered hydrophobic as a result of adsorption of surface contaminants.[121] Prakash et al. found a clear correlation between the increase in water contact angle and carbon coverage resulting from hydrocarbon adsorption, further corroborating that the observed hydrophobicity is not a reflection of the electronic structure of lanthanide cations.[122]

Appropriate chemical functionalization of a surface alters the surface free energy of a texturized membrane, making it selectively wettable toward permeation of either oil or water depending on the type of emulsion, thereby inducing separation at the surface. This is accomplished by preferentially diminishing adhesive interactions for water–oil with the surface while facilitating adhesive interactions (hydrogen bonding, ion–dipole interactions, and ion-induced dipole interactions) with oil–water, thereby resulting in differential wettability. Membranes functionalized with polymers have been extensively investigated over the last several decades; the majority are fabricated from poly(vinylidene difluoride) (PVDF),[123−126] polyurethane (PU),[111,127,128] polysulfone (PS),[113,129] polyacrylonitrile (PAN),[109,130] and polybenzimidazole (PBI).[110,131] Membrane configurations used for oil–water separations, including SAGD applications, are typically prepared by casting architected polymeric films[113,126] onto ceramic[124,125]/ceramic–polymer blend membrane supports using phase inversion.[126,127,132] Polymers such as polydopamine and poly(vinylidene difluoride) (PVDF) have been structured into precisely tunable porous networks, and furthermore, they can be used to modify the surface energy at solid–air and solid–liquid interfaces.[82,113] Polymer surface modification provides a means of sensitively tuning surface energies. Polymer swelling, as well as constraints on operational temperature,[133] represents the major drawbacks of polymeric membranes. Polymer swelling is the abrupt change in the volume of a polymer network by penetration of the solvent resulting from unbalanced osmotic and viscoelastic restoring forces.[134] This swelling can be resolved to some degree by reducing voids between polymer chains; polymer grafting with a functional organic side chain can further diminish polymer solubility in specific solvents, rendering the surface more resilient to solvent interactions.[110,135]

Graphene oxide materials have also found application in membrane design and display excellent hydrophilicity, attributable to the presence of carboxyl and hydroxyl groups on the edges of their basal plane that help create hydrogen bonds with water.[136,137] The incorporation of graphene oxide in the membrane is usually achieved by creating a dispersion with a polymer. Notable examples include reduced graphene oxide (r-GO)/graphitic carbon nitride (C3N4) dispersed in polydopamine (PDA),[138] graphene oxide dispersed in polybenizimidazole (PBI),[110] and graphene oxide bound to amine-terminated polyacrylonitrile (PAN) fiber,[109] where amide linkages are created between carboxyl groups on the graphene oxide and amine groups on the surface of PAN fibers. Graphene oxide incorporation increases the surface density of hydroxyls, thereby enhancing surface hydrophilicity.

Perfluorinated compounds are ubiquitously used to alter the wettability of surfaces to different liquids. The low polarizability of covalently bonded fluorine atoms in organofluorine compounds results in intrinsically low intermolecular forces, yielding a low interfacial tension.[139] However, creating a surface with low free energy reduces adhesive interactions for both oil and water,[117] which is not optimal for the separation of disparate liquids. Instead, fluorosurfactants can promote differential wettability. In such compounds, the fluorinated tail inhibits polar interactions with both oil and water, whereas the anionic head promotes polar interactions with water while exhibiting limited interactions with hydrocarbons. Such precise tunability of molecular interactions is imperative to create a selectively wettable membrane surface.[17,117,140,141] We have explored this facet of membrane development using potassium perfluorooctane sulfonate (PFOS) surfactant as a surface-energy modifier on a stainless-steel surface spray coated with tetrapodal ZnO to create mesoscale texturation and adhered with an amorphous silica network.[17] The membrane architectures exhibit excellent oil–water separation, decreasing oil content in water to <300 ppm in a single pass.

The adhesion of inorganic morphological architectures to the structural support membrane represents a substantial challenge for membrane development. Poor adhesion results in membrane degradation, fouling, and lower separation flux rates. Traditionally, organosilanes[82,125] have been used as precursors to create amorphous three-dimensional (3D) silica networks on texturized membrane surfaces through the Stöber process,[17,125] a sol–gel approach that involves hydrolysis of organosilanes, followed subsequently by condensation to form a siloxane network. Precise control of organosilane chemical activity is needed to achieve uniform surface distribution of the siloxane network on the membrane. Recent research has focused on using inorganic matrices created via condensation of aluminum phosphate[112,142] to bind textural elements on the surface of the membrane at elevated temperatures as per[143]Thermal curing constantly depletes water that is created as a product of condensation, thereby shifting the equilibrium in the forward direction, according to Le Chatelier’s principle. One disadvantage of this approach arises from the use of corrosive acids in high concentrations at elevated temperatures. An alternative approach has focused on exploiting the crosslinking of sodium alginate, a sodium salt of copolymer alginic acid that possesses carboxyl groups as repeating side-chain units.[123] These carboxyl groups form a coordination complex in the presence of divalent metal cations. Membrane surfaces are modulated to initiate complexation by imparting a net positive charge.[124,144] It is important to note that the techniques developed for improving adhesion are generally performed prior to chemical functionalization intended to modify surface energies and interfacial interactions at solid–liquid interfaces.

In past research, we have demonstrated the separation of SAGD viscous oil-rich emulsions exploiting the orthogonal wettability derived from the hierarchical texturation of ZnO tetrapods arrayed onto a stainless-steel mesh.[16] An amorphous silica layer was deposited to tether ZnO tetrapods onto stainless-steel through the formation of siloxane linkages (Figure A–C). The 3D textured membranes show a distinctive combination of oleophilic and superhydrophobic behavior. Membrane separation was achieved within a custom-designed thermal autoclave in the presence of endogenously generated pressure. At operational temperatures >130 °C, complex emulsions were disrupted, reducing water content in the permeate to 0.69 vol % in a single pass. A related approach by Kollarigowda et al.[108] focused on the use of finely textured block copolymers applied to a cellulose fiber base, which selectively eliminates thiol contaminants along with affecting oil–water emulsion separation. A 3-(trimethoxysilyl)propyl acrylate-block-myrcene block copolymer was chemically attached to the cellulose membrane using reversible addition-fragmentation chain transfer polymerization (Figure D–F). The grafting of the block copolymer reversed the wettability behavior of the cellulosic membrane from superhydrophilic to superhydrophobic and oleophilic. Myrcene moieties on the membrane surface reacted with thiol contaminants, through “thiol-ene” chemistry,[145] eliminating noxious odors from the separated water.

Figure 6

(A, B, D, E, J, K) Scanning electron micrographs depicting different membrane supports and surface morphology modifiers. (A) Top view of stainless-steel membranes functionalized with ZnO tetrapods and tetraethoxysilane (TOES)[16] and (B) individual ZnO tetrapods used for imparting mesotexturation at the stainless-steel surface. Reprinted (adapted) with permission from O’Loughlin, T. E.; Ngamassi, F.-E.; McKay, P.; Banerjee, S., Separation of Viscous Oil Emulsions Using Three-Dimensional Nanotetrapodal ZnO Membranes. Energy and Fuels 2018, 32 (4), 4894–4902. Copyright 2018 American Chemical Society.[16] (C) Chemical structure of tetraethoxysilane; (D) top view of the dehydrated dermis of the cellulose membrane functionalized with 3-(trimethoxysilyl)propyl acrylate-block-myrcene[146] and (E) nanoscale texturation imparted on the block copolymer-functionalized cellulose membrane (inset micrographs are at higher magnification). Reprinted (adapted) with permission from Kollarigowda, R. H.; Abraham, S.; Montemagno, C. D., Antifouling cellulose hybrid biomembrane for effective oil–water separation. ACS applied materials and interfaces 2017, 9 (35), 29812–29819. Copyright 2017 American Chemical Society;[146] (F) chemical structure of 3-(trimethoxysilyl)propyl acrylate-block-myrcene; (G) side view of polyethersulfone/TiO2 nanotubes membrane;[113] (H) top view of individual TiO2 nanotubes used to create nanocomposite membranes, licensed with permission under CC BY 4.0;[113] and (I) chemical structure of polyethersulfone (PES) and polyvinylpyrrolidone (PVP).

The cleaning of SAGD wastewater has been recently demonstrated by Mahdi et al.,[113] who used a hybrid membrane incorporating hollow TiO2 nanotubes and polyethersulfone. The overall polymer matrix possessed a mean pore radius of ca. 4 nm that varied with the TiO2 nanotube loading. TiO2 nanotubes (average height of 10 μm and a diameter of 250 nm), polyethersulfone (PES), and polyvinylpyrrolidone (PVP) were dispersed in dimethylacetamide solvent and cast onto a glass surface (Figure G–I). Next, nonsolvent-induced phase separation in a coagulation bath of deionized water was used to prepare a superhydrophilic membrane. At 1 wt % loading of TiO2 nanotubes, hybrid membranes removed 53.9% of dissolved organic matter from SAGD-produced water and showed a relatively low predilection for fouling.

Attallah et al.[82] have designed poly(ethylene oxide) (PEO)-based membranes incorporating a γ-Al2O3/TiO2 ceramic support tethered using organosilanes. These membranes have effective pore dimensions less than 10 nm and were successful in reducing the organic carbon content of the separated water by approximately 91%, along with a 47% reduction in sulfates. Modified membranes with a γ-Al2O3 base support exhibit flux rates in the range of 50–150 L/(h m2) that increase over time, significantly greater than their unmodified counterpart. In contrast, membranes modified with a TiO2 base support exhibited higher flux rates, ranging 220–350 L/(h m2) that slightly decreased over time.

The effective design of membranes for the separation of SAGD emulsions requires a deeper understanding of the nature of emulsions and their interactions with membrane surfaces. Mechanistic understanding of interfacial interactions and the development of systematic composition—mesoscale-structure–function correlations—will enable rational design of orthogonally wettable membrane architectures as well as the selection of process conditions that aid in the destabilization of emulsions. The design of membrane architectures and processes that mitigate fouling and yield high flux rates is particularly important to enable their integration within modular mobile media filtration units that are being increasingly adopted and aim to process over 10 000 barrels a day per unit.

Membrane-Assisted Removal of Diluent during Oil Refinement in SAGD

As discussed in the preceding section, a wide variety of materials have been used to prepare membranes for the separation of oil–water emulsions based on differences in wettability. However, the selective separation of different organic molecules remains less developed. A particular problem pertains to the separation of diluents, lighter hydrocarbons added to modify the flow properties of bitumen. As noted in the introductory sections, a substantial portion of pipeline capacity is consumed in the flow of diluent back and forth between production sites and refineries in North America.[147] Globally, ca. 100 million barrels per day of oil are transported to refineries, with unconventional heavy oil that requires dilution claiming an ever-increasing share.[148] The most commonly used diluents are hydrocarbon solvents such as naphtha, kerosene, light natural gas condensates, and light crude oils.[147,149−151] To achieve typical pipeline specifications of a density not to surpass 940 kg/m3 and a viscosity not to surpass 350 cSt, ca. 25–50% (v/v) of light hydrocarbons are required.[147,149,152−154] Distillation processes used to separate diluents at refineries based on differences in molecular weight and boiling point are exceedingly energy intensive.[155−157] Distillation, in general, consumes 1100 TWh per year, representing ca. 1% of global energy consumption.[148] Several novel technologies have been proposed to optimize the functioning of distillation units to reduce energy consumption without alterations in infrastructure leveraging advances in energy analysis, nonlinear programming, simulation tools, and mixed-integer nonlinear programming.[155,156,158,159] These modeling techniques have enabled the parametric optimization for distillation units to achieve enhancements in energy efficiency and heat recovery.

The availability of energy-efficient nonthermal methods such as membrane-based technologies to separate diluents from heavier bitumen fractions would transformatively alter the economics of solvent recovery with substantial implications for the viability of the SAGD oil recovery process.[15,117] As inspiration and rationalization for membrane design to separate dilbit, reverse osmosis membranes are widely used for water desalination and have extensively supplanted thermal desalination. Approximately 320 000 m3 of clean water per day is produced by reverse osmosis, almost twice the volume of crude oil handled by refineries.[160,161] Using membranes to separate mixtures of hydrocarbons is an emerging area that promises decreased operational time and reduced energy consumption; the latter derives from mitigating the liquid-to-vapor phase change required in conventional separation and purification procedures. Membrane methods can potentially fractionate highly complex mixtures exploiting differences in molecular size, shape, and surface interactions and are anticipated to bring about as much as a 10× increase in energy efficiency as compared to current energy-intensive thermal methods.[148]

One of the biggest challenges for membrane-based technologies is the separation of molecules with similar kinetic diameters with selective permeability for one type of organic molecule over another. Pore size exclusion and sorption–diffusion are the main mechanisms for liquid-phase separations through membranes.[162,163] In size exclusion, the permeate is a liquid phase that travels through the pores with solute species that are below the molecular weight cutoff of the membrane. In sorption–diffusion separations, the feed as a liquid phase is partially vaporized into the membrane surface promoted by a difference in chemical potential, and that results in the diffusion of the components of the permeate into the membrane, and consequently, desorption of the permeate in the vapor phase at the downstream side of the membrane takes place. Castro-Muñoz et al. note that the efficacy of separation in membrane-based technologies depends on several factors encompassing the composition of the mixture sought to be separated, operational parameters, and distinctive membrane features.[164] Specifically, it is essential to consider the type of solvent/solute, molecular weight of solute/solvent, polarity, and solute charge for enhancing efficiency.[162,164] Furthermore, parameters such as flow rate, temperature, and pressure are also characteristics to consider for parametric optimization in the separation of different components of a solution.[164−166] In terms of the membrane design, it is important to evaluate the membrane material, the architecture and flow geometry of the module, and characteristics such as pore size, wettability, tortuosity, and porosity.[164−166]

A balance between permeability and selectivity is pivotal in achieving optimal separation. Typically, an enhancement in permeability is inversely correlated with the selectivity of permeation. The pore size distribution represents a key parameter that has to be optimized to ensure optimal separation performance.[162,167,168] A variety of materials have been used to modulate pore dimensions to achieve the selective permeability of a membrane for liquid-phase separations. Some of these materials that can potentially engender selective separation of the mixtures of organic molecules include carbon molecular sieves (CMS), diamondlike carbon (DLC), graphene oxide (GO), crosslinked microporous polymers, linear polymers, polymers of intrinsic microporosity (PIMs), crosslinked polymers with contorted monomers, conjugated microporous polymers (CMPs), metal–organic frameworks (MOFs), covalent organic frameworks (COFs), porous organic cages (POCs), and zeolites.[160,162] Of these materials, COFs are covalently interconnected crystalline porous networks with tunable pore dimensions built from multifunctional structural units with substantial scope for functionalization of pore edges. Pore tuning in COFs is achieved based on engineering the length and structure of linkers, anchoring a variety of functional groups onto the pore walls of the framework, and through alterations of the stacking modes of two-dimensional (2D) frameworks.[169,170] As an illustrative example, Dey et al. synthesized a variety of COF-based membranes through interfacial crystallization and accessed pore dimensions ranging from 1.4 to 2.6 nm.[170,171] These modifications were achieved by the incorporation of a series of linear linkers that were combined with a framework of 1,3,5-triformylphloroglucinol (Tp) trigonal planar linkers to assemble a pore geometry that allowed for the selective permeation of small molecules such as acetonitrile, water, ethanol, and methanol.

Tuning the chemistry of pore surfaces is key to achieving liquid separations. Nagai et al. tailored the pore dimensions of COFs synthesized by condensing hexahydroxytriphenylene (HHTP) with azide-appended benzene diboronic acid (N3-BDBA) and benzene diboronic acid (BDBA). The azide units enable the use of click chemistry reactions with alkynes to generate 2-propynyl acetate on the wall of the framework, enabling reduction of pore dimensions from ca. 3.0 to 1.2 nm. In another example, Li et al. prepared nanoporous COF-based membranes by modified interfacial polymerization through the incorporation of phenolic hydroxyl groups on triformylbenzene monomers, which modified the stacking mode of the 2D COF framework from AA to AB stacking, thereby providing access to highly constrained ca. 0.6 nm channels with a uniform pore distribution.[172] This pore geometry showed an excellent ability to exclude relatively high-molecular-weight compounds such as 1-butanol and phenylcarbinol while permitting permeation of smaller molecules such as water and methanol. As such, postsynthetic modification of COFs holds considerable promise for tuning pore dimensions and enhancing membrane selectivity.[173]

MOF-based membranes for liquid-phase separation have been widely used in different fields such as aromatic separations, liquid fuel purification, water treatment, gas separations, and solvent recovery.[174,175] These crystalline porous networks exhibit high porosity, high surface area, tunable pore dimensions, scope for introducing flexibility, and thermal stability.[176] In terms of engendering separations, the key features that make these materials attractive include pore dimensions and geometries that are readily tunable based on the selection of inorganic nodes and organic linkers; furthermore, pore edges can be rendered amenable to functionalization.[175] MOF-based separations typically exploit preferential adsorption and shape selectivity. MOF-based membranes have found extensive use in gas-phase separations of hydrocarbons; liquid-phase separations are relatively underexplored.[174,175] Nevertheless, Diestel et al. developed a polycrystalline MOF membrane utilizing a zeolitic imidazolate framework (ZIF-8) to separate liquid mixtures of n-hexane with benzene and n-hexane with 1,3,5-trimethylbenzene (mesitylene) through pervaporation.[174,175,177] Based on previous studies on ZIF-8 membranes, bulky alkanes and aromatics are not capable of migrating through the narrow pores of the membranes, while linear alkanes are adsorbed due to their characteristic sieve exclusion behavior.[175,177] The experiments illustrated that the single components, n-hexane, and benzene were adsorbed onto the surface of the ZIF-8 membrane, whereas mesitylene was not adsorbed to the porous ZIF-8 framework. In the separation of n-hexanes and benzene mixtures, the more mobile component of the mixture, n-hexane, was blocked by the less mobile bulky molecule, benzene. When considering the mixture of n-hexane and mesitylene, a molecular sieving phenomenon was observed, and the flux of n-hexane was dependent on the concentration of mesitylene. Therefore, for the mesitylene to be separated from n-hexanes by the porous network, the molar mesitylene ratio in the mixture should be less than 50% since a high concentration of mesitylene can block the pore entrance for the mobile component (n-hexane).

Surface modification of MOF-based membranes has rendered a beneficial outcome in the separation and recovery of aromatic compounds, which is pivotal in the oil and gas industry. For instance, Zhao et al. developed a series of functionalized metal–organic polyhedral (MOP)/hyperbranched polymer hybrid membranes for the recovery of aromatic hydrocarbons from aromatic–aliphatic mixtures.[178] They evaluated the effects of functional groups on MOF-based membranes architectures on the separation of aromatics–aliphatic mixtures in terms of adsorption of aromatics. A hybrid MOP-SO3NaH membrane for the separation of toluene and n-heptane exhibited a permeation flux of 528 g/(m2 h) for toluene recovery, whereas a benzene and cyclohexane mixture revealed a flux rate of 540 g/(m2 h) for benzene recovery. The separation of benzene and cyclohexane is currently one of the most challenging separations in the chemical industry due to their indistinguishable boiling points, ca. 0.6 °C of variation.[179] The presence of a sulfonate group on the modified MOP membrane enhanced the adsorption selectivity of toluene over n-heptane and of benzene over cyclohexane due to the π–π interactions between the sulfone-conjugated MOP framework and toluene. It has been demonstrated previously that polymers functionalized with a strong electron-withdrawing group, such as sulfone, can potentially interact with the π-electron clouds of aromatic hydrocarbons to increase the selectivity and permeability at the surface.[178,180] The permeability and selectivity of these hybrid membranes were dependent on functional groups present on the surface of the framework. The MOF structures with functional groups with high polarity promoted a feasible adsorption or recovery of aromatic hydrocarbons.

Polymer-based membranes offer a combination of facile processability, flexibility, and scalability for the separation of liquid mixtures. As an illustrative example, Thompson et al. developed a highly thermally stable N-aryl-linked spirocyclic polymer membrane, which exhibited enrichment of molecules lighter than 170 Da, enabling the selective permeation of hydrocarbons with less than 12 carbon atoms and boiling points less than 200 °C (Figure ).[148] This spirobifluorene-aryl-diamine-based membranes exhibited a mechanically robust surface with a narrow distribution of ultramicropores as a result of efficient chain packing designed through the incorporation of flexible C–N linkages that promote π–π stacking interactions at the surface. The membrane architectures enabled the separation of jet fuel and gasoline from a crude oil mixture. Thomson and co-workers suggest that the separation was engendered due to the presence of aromatic C–N linkages that provide a combination of hindered rotation and optimal pore distribution to promote the transport of molecules through the micropores at rates governed by molecular size and dimensions.

Figure 7

Fractioning of light crude oil mixtures using a spirofluorene aryl diamine polymer-based membrane exhibiting microporosity. (A) Chemical composition of the rigid polymeric spirofluorene aryl diamide series showing a variety of aromatic biphenyl–phenyl linkages that were evaluated for enhancing the separation efficiency of organic molecules based on molecular size and type. (B) Comparison of boiling-point distributions of the feed, permeate, and retentate from the fractioning of crude oil using a spirocyclic membrane backbone with aromatic diamine I, which exhibits a narrow pore diameter distribution. The permeate exhibits notable enrichment of lighter hydrocarbon molecules. (C) Gas chromatography-flame-ionization detection (GC-FID) analysis of the membrane fractioning of crude oil, which correlates the feed and permeate solutions based on boiling point, polarity, and response factor. The enrichment of molecules lighter than the carbon number of 12 was noted. From Thompson, K. A.; Mathias, R.; Kim, D.; Kim, J.; Rangnekar, N.; Johnson, J. R.; Hoy, S. J.; Bechis, I.; Tarzia, A.; Jelfs, K. E.; McCool, B. A.; Livingston, A. G.; Lively, R. P.; Finn, M. G., N-Aryl-Linked Spirocyclic Polymers for Membrane Separations of Complex Hydrocarbon Mixtures. Science 2020, 369 (6501), 310–315. Reprinted with permission from AAAS.[148]

Currently, developed membrane-based methods used for the separation of organic solvent mixtures include pervaporation, nanofiltration, reverse osmosis, and forward osmosis.[160,162] Typical membrane systems such as microfiltration, ultrafiltration, nanofiltration, and reverse osmosis exploit distinctive pore size diameters ranging from 100 to 10 000 nm, 2 to 100 nm, 0.5 to 2 nm, and 0.1 to 1 nm, respectively.[164,181] Nanofiltration membranes designed to sieve organic molecules need to selectively permeate molecules in the range of 200–1000 Da, contain pores below 2 nm, and exhibit chemical stability across a wide variety of solvents.[162] Indeed, organic solvent reverse osmosis membranes can sieve organic molecules with a molecular weight lower than 100 Da.[182] Some of these methods have been scaled to industrial operations. For instance, nanofiltration membranes are currently employed at ExxonMobil’s Beaumont refinery to produce lube oil and as a part of dewaxing operations.[160,183] These polymeric membranes are constituted from polyimide and are used to selectively segregate lube oil and recover a mixture of ketones or aromatics including methyl ethyl ketone, methyl isobutyl ketone, and toluene.[183] As another example, modules developed by ExxonMobil Corporation and W.R. Grace have been used to separate linear alkanes with a molecular weight less than 300 Da from lube oil molecules to successfully reduce fuel oil consumption, greenhouse gas emissions, water usage, and emissions of volatile organic materials.[183]

Ideally, membranes for separating different hydrocarbons by properties such as molecular weight or boiling point should exhibit high permeability for one phase concurrent with high rejection above a threshold molecular weight or kinetic diameter. In addition, the membranes should exhibit a combination of mechanical stability, entropic selectivity, and scalability. By taking advantage of current developed concepts and technologies for liquid–liquid, liquid–solid, and gas-phase separations, membrane systems hold promise for enabling effective separation of diluent from bitumen as well as different fractions of bitumen, thereby alleviating the considerable energy and cost burden of conventional thermal distillation processes.

Corrosion

Corrosion is the gradual degradation of a material as a result of chemical reactions with the surrounding environment (Figure ).[14] While corrosion represents a considerable drag on resources, in the case of base metals, it is inevitability driven by strong thermodynamic driving forces. The oxidation of base metals tends to be strongly exothermic as a result of high enthalpies of metal–oxygen bonds. As such, protecting base metals essentially comes down to delaying the kinetics of corrosion reactions by establishing diffusion limitations at metal interfaces, effectuating electrochemical polarization to increase overpotentials for corrosion reactions, selectively triggering the formation of passivation layers at interfaces, or triggering sacrificial reactions that spare the components sought to be protected.[184] Throughout both upstream and midstream processes, corrosion is a challenge across the entire infrastructure spanning from the wellhead to the refinery. The SAGD infrastructure is primarily constructed from low alloy steels as a result of performance constraints and cost considerations and is subject to flow-accelerated corrosion, localized corrosion, erosion, and erosion–corrosion.[25] When looking at pipeline production failures in Alberta, Canada, from 1980 to 2005, 58% failed as a result of internal corrosion and 12% from external corrosion.[25] About 25% of transmission pipeline accidents are caused by corrosion.[25] Extensive studies have investigated fracture patterns (intergranular vs transgranular) that result from stress corrosion cracking and have identified origins as being concurrent potent environments (ranging from pH 6.5–9), use of susceptible materials, and inhomogeneous stress gradients.[185,186] Much research has been invested in advancing mechanistic understanding so that even if corrosion cannot be entirely prevented, it can be substantially slowed down and detected before catastrophic failure. Methods to inhibit corrosion can be clubbed into five broad categories: barrier protection, cathodic protection, anodic passivation, active corrosion inhibition, and “self-healing.”[14]

Figure 8

(A) Schematic illustration of corrosion at the surfaces of base metals. (B) Schematic depiction of the tortuous path when at low graphene concentration and percolative network formed with high graphene concentration. Reprinted (adapted) with permission from Davidson, R. D.; Cubides, Y.; Fincher, C.; Stein, P.; McLain, C.; Xu, B.-X.; Pharr, M.; Castaneda, H.; Banerjee, S., Tortuosity but not Percolation: Design of Exfoliated Graphite Nanocomposite Coatings for Extended Corrosion Protection of Aluminum Alloys. ACS Applied Nano Materials 2019, 2 (5), 3100–3116. Copyright 2019 American Chemical Society.[194] (C) Corrosion inhibition with tungsten carbide near-nanocrystalline cladding on steel. Reprinted from Corrosion Science, 53/6, Saha, G.; Khan, T.; Zhang, G., Erosion–Corrosion Resistance of Microcrystalline and Near-Nanocrystalline WC–17Co High Velocity Oxy-Fuel Thermal Spray Coatings, 2106–2114, 2011, with permission from Elsevier.[199]

Barrier protection provides an impermeable layer that prevents corrosive species from reaching the substrate, thereby increasing the effective tortuosity. Cathodic protection uses a sacrificial anode that polarizes the substrate, such that the anode oxidizes rapidly before the substrate. Anodic protection uses the formation of a passivating, impermeable layer to impede the redox reaction. Active corrosion inhibition uses strongly adsorbing Lewis bases that are released upon damage to form a protective barrier at exposed interfaces. “Self-healing” mechanisms involve reconstitution of coatings from the release of monomeric precursors from micro- and nanocontainers or encapsulated particles. As an alternative strategy, corrosion is mitigated by supplanting low alloy steels in components such as piping, valves, fittings, and vessels with cement, plastics, or nanocomposites. For instance, cement pipes can then be internally coated with plastic or fiberglass to reduce friction; however, such tubing is plagued by lack of ductility and exhibits a tendency to undergo plastic deformation as a result of stresses developed during thermal cycling.[25] Plastic piping (polyester resin) is a viable alternative and affords corrosion resistance but is typically constrained in its ability to handle high pressures and temperatures.[25] Therefore, utilizing mechanisms of corrosion control is paramount to viable and efficient production within the oil sands.

In lease tanks, trucks, and rail cars, corrosion is likely to occur along the vessel floor as a result of the accumulation of water; the high sulfur content (sourness) of most heavy oil increases the susceptibility of midstream infrastructure to corrosion, as internal corrosion by acidic gases such as CO2 and H2S are the primary origins of failure.[25] These gases can further precipitate hydrogen embrittlement, stepwise cracking, and hydrogen blistering.[25] Currently, industry-standard coatings are zinc alloys (often deposited by methods such as hot dip galvanization) that provide sacrificial cathodic protection.[187,188] The extent of corrosion inhibition is directly proportional to the thickness of zinc alloy. Zinc alloy coatings are prone to flaking and cracking as a result of the less ductile nature of the coatings in comparison to underlying steel substrates. Chromium and other chromate coatings have been extensively explored as a means of corrosion inhibition as a result of their outstanding wear resistance, ease of plating, and corrosion resistance. The outstanding corrosion inhibition afforded by such coatings derives from the facile formation of a passivating layer triggered by changes in local pH upon initiation of corrosion and enabled by the extremely low-solubility products of chromate salts.[189,190] However, extensive regulations resulting from the presumed carcinogenicity of chromium have severely restricted the use of hexavalent chromium; chromium concentrations in air, water, and soil are actively monitored and subject to stringent regulations across the world.[187]

Alternatives to zinc and chromate coatings include polymers, metal oxides, and self-assembled molecular monolayers.[191−193] Substantial research has focused on the development of nanocomposite thin-film coatings that incorporate nanoparticles in various base polymer matrices to activate multiple modes of corrosion inhibition. For instance, we have designed hybrid nanocomposite coatings that incorporate multiwalled carbon nanotubes or graphene with a polyetherimide (PEI) matrix to offer synergistic corrosion inhibition to low alloy steels through an “active–passive” mechanism.[187] PEI shows excellent adhesion to steel (as well as to Al alloys) and can be cast as pinhole-free films that afford considerable barrier protection. Tafel plot analysis shows the formation of an expanded passivation window upon the incorporation of graphene, translating to approximately 3 orders of magnitude diminution of corrosion rates.[187] The loading of graphene and its connectivity within the nanocomposite is crucial to corrosion inhibition. A systematic study of different graphene concentrations in PEI has identified two distinct regimes, below and above percolation thresholds for graphene.[194] At low graphene concentrations, graphene fillers induce greatly increased tortuosity of ion diffusion pathways, enabling excellent corrosion protection of underlying aluminum alloys, whereas at high concentrations, graphene flakes form a percolative network and initiate deleterious galvanic corrosion (Figure C).[194] As an alternative strategy, highly electroactive Mg nano and microparticles have been embedded within PEI and epoxy/polyamide matrices, wherein they afford sacrificial cathodic protection.[192,195−198]

In a SAGD operation, tailing pipelines used for the transportation of the slurry, containing water, clay, silt, and sand that remains after the heavy oil is recovered from the oil sands are especially prone to degradation. Tailing pipelines suffer from both erosion and corrosion that is a result of synergistic mechanical abrasion and electrochemical processes.[25,199,200] Solutions to prevent corrosion and erosion within pipelines involve cladding with ceramic or metal composites, polymer linings, and surface modification. Success has been observed with ceramic materials (TiO2, SiC, and diamond with a Ni–P alloy electroless deposition) as they afford a combination of high hardness and chemical stability.[201,202] For instance, Neville et al. have reported a tungsten carbide metal matrix composite cladding, deposited using plasma-transferred arc welding, which provides a substantial increase in corrosion resistance. The combination of hard tungsten carbide particles with the ductile metal matrix binder (Ni, Cr, Si, B, and Fe) allows for increased strength and fracture toughness of the composite.[199] Saha et al. have also explored a tungsten carbide coating incorporating cobalt deposited using high-velocity oxy-fuel process and compared the corrosion potential as a function of coating microstructure, contrasting microcrystalline and nanocrystalline coatings with pristine steel (Figure C).[201] They observed that the difference in microcrystalline and nanocrystalline coatings altered the dominant failure mechanism, in that the former failed due to erosion and the latter from corrosion. The nanocrystalline coating had a reduced erosion–corrosion rate by one-third compared to the microcrystalline coating.

The viability of heavy oil production from unconventional deposits depends not just on efficacy and cost but also on minimizing environmental impact. As such, the close monitoring of infrastructure components and ensuring their integrity is of paramount importance. Preventing failures resulting from corrosion is at the heart of this puzzle and requires integrated corrosion control systems. Incorporating coatings that exhibit multiple modes of corrosion inhibition offers the greatest defense against the synergistic mechanisms of deterioration. The development of efficient membranes that allow for the cleaning of produced water and heavy oil, as described in Section 3 will yield cleaner oil and water streams, is thus integrally linked to the preservation of midstream infrastructure. Considerable research is focused on active corrosion inhibition and self-healing coatings as a means of restoring coating integrity upon damage as well as in improved real-time diagnostics of corrosion combining electrochemical measurements and other nondestructive evaluation tools with machine learning.[203−205]

Solutions for Midstream Transportation

Coatings

Crude oils with API gravities <22.3° (specific gravities> 920 kg/m3) are generally classified as being heavy oils. Extra heavy crudes have API gravities <10 and specific gravities in excess of 1000 kg/m3. The largest flow streams in Canada, Western Canadian Select and Access Western, have an API gravity range from 18 to 22; these streams further have sulfur content in excess of 3.5%. In comparison, Western Texas Intermediate has an API gravity of ca. 40 and a sulfur content of <0.5%. The transportation of heavy crude oils by pipeline, railcar, truck, or tanker gives rise to a substantial set of challenges stemming from the high viscosity of these fuels; for instance, bitumen has a viscosity exceeding 300 000 cP and thus requires extensive thermal jacketing or dilution with lighter hydrocarbons to induce flow. In addition, bitumen deposits in the Athabasca region of North America are landlocked, precluding the use of large tankers. As such, there is considerable reliance on pipeline transport (Figure ), albeit pipeline capacities are limited in every direction and unable to transport the full production of the Canadian Oil Sands to refineries.

For pipeline transport of heavy crude oil, rheological modifiers such as diluents, surfactants, polymer additives, and emulsifiers have been extensively used to reduce the viscosity, alter flow profiles, and decrease frictional forces such as to enable transportation of these fluids at reasonable pumping powers.[206−208] However, a major drawback to these approaches is that such additives must then be removed and pumped back to the point of origin; much valuable pipeline capacity is consumed in the flow of light condensates added solely for the purposes of rheology modification, which furthermore necessitate a substantial infrastructure for solvent recovery at terminals. As an entirely different approach, much recent effort has focused on the modification of the interior surfaces of pipelines to alter flow profiles. By reducing frictional forces at the solid–liquid interface, the overall drag in the system can be reduced, which in turn reduces the amount of diluent required and lowers the pumping power necessary to flow heavy crude oil within a pipeline. Such surface modification further reduces the extent of fouling of the solid surface in pipelines, bitutainers, rail cars, and shipping vessels, thereby increasing the efficiency of recovery. In this section, we will focus on surface coatings capable of facilitating drag reduction as well as alternative strategies for partial bitumen solidification to enable solid-phase transport.

To decrease drag in a system through direct modification of the solid surface, localized regions of slip must be formed to break the no-slip boundary condition in laminar flow and to reduce solid–liquid interactions under turbulent flow (Figure B).[209−216] While surface energy (measured in the form of contact angles) plays a substantial role in governing the formation of slip planes, the 3D texture of the surface is of paramount importance and determines the formation and interconnectedness of trapped air pockets known as plastrons.[15,117,209,216−218] Such plastrons provide regions of localized slip, as depicted in Figure B. According to Young’s equation (eq ), the surface roughness (r) acts as a scaling factor that modifies the inherent wettability of a surface, which is determined by the interplay between the solid–liquid (γSL), liquid–vapor (γLV), and solid–vapor (γSV) interfacial energies.[219,220] As a result of the high surface tension of many liquids, the interfacial energy of the solid–liquid interface must be substantially reduced to prevent wetting of the surface. Plastrons resist fluid permeation and function as localized regions of slip since friction between the liquid and air pocket is very low. However, it is worth noting that the plastronic surface is metastable; beyond a specific breakthrough, the oil will permeate into the pores and establish a Wenzel wetting regime.

Figure 9

(A) Schematic illustration of a textured surface with a standard post structure, a reentrant structure, and a doubly reentrant structure, where γ is the surface tension and Δp is the differential pressure. (B) Schematic diagram of a representative velocity profile on a smooth surface and a textured omniphobic surface. Copyright 2020 Wiley. Used with permission from (Cool, N.; Douglas, L.; Gupta, S.; Banerjee, S., Hierarchically Textured Oleophobic Internal Coatings that Facilitate Drag Reduction of Viscous Oils in Macroscopic Laminar Flow. Advanced Engineering Materials).[209]

While the modified Young’s equation suggests that the surface roughness acts solely as a scaling factor to enhance the inherent wettability of a surface, upon closer examination of the force balance between the gravitational forces acting on the test liquid and the direction of the surface tension, it becomes apparent that the details of the geometry of the plastron play a determining role.[216,217,221] Considering a straight pillarlike geometry (Figure A), the inherent contact angle must remain above 90° for a liquid to be repelled by the surface. However, when the plastron opening is characterized by the reentrant curvature as sketched in Figure A, a liquid droplet can be repelled by the surface even when the intrinsic contact angle is below 90°. This is a direct result of the force applied on the droplet by surface tension, γ, being oriented opposite in direction to the differential pressure, Δp.[217,221] As the inherent contact angle approaches 0°, the surface tension of the liquid is directed parallel to the surface with the vertical component being oriented opposite in direction to the applied pressure. As such, a doubly reentrant surface curvature is required to repel a completely wetting liquid with an inherent contact angle of 0°.[216] Given a doubly reentrant surface opening, a completely wetting liquid spreads along the surface of the material and begins to wrap around the edges of the post; however, since a significant projection of the surface tension is oriented vertically in opposition to the direction of the applied pressure, it is able to repel the fluid from further permeating within the plastron. It is worth noting that such a scenario requires a specific ratio of topographical features with the appropriate curvature to fully repel the liquid. If there are only a sparse few topographical elements, liquid droplets are much more readily able to outweigh the breakthrough pressure; in such a scenario, the net vertical force derived from the surface tension of the liquid is readily surpassed by the positive pressure from the liquid. Analogously, if topographical units are in close proximity, the ratio of liquid in contact with the solid surface greatly exceeds the liquid in contact with air within the plastrons, once again enabling the liquid to wet the surface. As such, a stringent set of criteria must be satisfied with respect to the shape of the plastrons, their areal density, and the relative spacing between topographical features to fully repel a liquid.[216]

A number of key descriptors have been discovered that can effectively bring about drag reduction and antifouling properties through the reduction of friction at the solid–liquid interface. Liu et al. made use of reactive ion etching (RIE) to form doubly reentrant structures of SiO2 and used these precisely patterned surfaces to confirm the importance of geometry, areal density, and spacing of textural elements on omniphobicity (Figure A).[216] Not only do the authors demonstrate that the doubly reentrant SiO2 surface is unwettable by water (with a surface tension of 72.8 mN/m) without any additional chemical functionalization, they furthermore demonstrated that it is not wetted even by perfluorohexane (C6F14, 3M Fluorinert (FC-72)), which has the lowest known surface tension of all liquids at 10 mN/m.[216] These results illustrate the paramount role of texture in the design of omniphobic surfaces for drag reduction.[222−225] However, lithographic patterning of doubly reentrant architectures has limited viability for applications at the scale of midstream infrastructure, and thus efforts have focused on striking a balance between textural elements that stabilize plastrons and chemical functionalization to reduce surface energy.

Figure 10

(A) Doubly reentrant surface. From Liu, T. L.; Kim, C.-J. C., Turning a Surface Superrepellent Even to Completely Wetting Liquids. Science 2014, 346 (6213), 1096–1100. Reprinted with permission from AAAS.[216] (B) Reentrant surface formed through the electroless deposition of nickel and PTFE beads. Copyright 2020 Wiley. Used with permission from (Cool, N.; Douglas, L.; Gupta, S.; Banerjee, S., Hierarchically Textured Oleophobic Internal Coatings that Facilitate Drag Reduction of Viscous Oils in Macroscopic Laminar Flow. Advanced Engineering Materials).[209] (C) Reentrant surface formed through the deposition of ZnO tetrapods on a steel mesh. Reprinted (adapted) with permission from O’Loughlin, T. E.; Dennis, R. V.; Fleer, N. A.; Alivio, T. E. G.; Ruus, S.; Wood, J.; Gupta, S.; Banerjee, S., Biomimetic Plastronic Surfaces for Handling of Viscous Oil. Energy and Fuels 2017, 31 (9), 9337–9344. Copyright 2017 American Chemical Society.[117] (D) Reentrant surface formed through the templating of TiO2 nanospheres using polystyrene beads. Reprinted (adapted) with permission from Douglas, L. D.; O’Loughlin, T. E.; Chalker, C. J.; Cool, N.; Gupta, S.; Batteas, J. D.; Banerjee, S., Three-Dimensional Inverse Opal TiO2 Coatings to Enable the Gliding of Viscous Oils. Energy & Fuels 2020, 34 (11), 13606–13613. Copyright 2020 American Chemical Society.[218]

An approach to define reentrant curvature at macroscopic scales inaccessible through precision patterning methods such as electron beam lithography, focused ion beam deposition, and ultraviolet lithography[226] that has attracted much recent attention is nanoimprint lithography. In this approach, a precise template with micro- and nanostructured roughness is prepared and the imprint is mechanically transferred onto the substrate. The compatibility of nanoimprint lithography with roll-to-roll and roll-to-plate processing holds promise for large-scale applications (Figure A).[227] Ye et al. have demonstrated that this technique can be applied with sub-30 nm resolution.[228] Choi et al. developed an oleophobic coating by combining reverse nanoimprint lithography with reactive ion etching using a polydimethylsiloxane template; the patterned surfaces were functionalized with a fluoroalkylsilane, specifically, heptadecafluoro-1,1,2,2-tetrahydrodecyl trichlorosilane (HDFS), self-assembled monolayer.[229]Figure B shows the precise control afforded by this method in the imprinting of various geometries, including arrays of cones, pillars, holes, and lines within the overhang structures.[229] Choo et al. demonstrated the use of UV-nanoimprint lithography in the precise patterning of perfluoropolyether surfaces; these authors used UV-cured polyurethane acrylate to replicate the complex texture of rose petals (Figure C).[230] Such surfaces emulate the detailed hierarchical texture of rose petals including micropapillae and nanofolds; the wettability of rose petals toward water droplets is further captured in high fidelity in these intricately patterned architectures.

Figure 11

(A) Schematic depiction of the roll-to-plate nanoimprint lithography technique, licensed with permission under CC BY 4.0.[227] (B) UV-nanoimprint lithography with control of geometric shape: (a) cone, (b) pillar, (c) hole, and (d) line. Reprinted (adapted) with permission from Choi, H.-J.; Choo, S.; Shin, J.-H.; Kim, K.-I.; Lee, H., Fabrication of Superhydrophobic and Oleophobic Surfaces with Overhang Structure by Reverse Nanoimprint Lithography. The Journal of Physical Chemistry C 2013, 117 (46), 24354–24359. Copyright 2013 American Chemical Society.[229] (C) Panels (a) and (b) depict rose-petal texturation and (c) and (d) depict polymeric templated rose petals. Reprinted from Materials Letters, 121, Choo, S.; Choi, H.-J.; Lee, H., Replication of Rose-Petal Surface Structure Using UV-Nanoimprint Lithography, 170–173, 2014 with permission from Elsevier.[230]

A nonlithographic approach for constructing reentrant surfaces at larger scales has been developed through modification of metal meshes through techniques such as electrospinning, candle soot coating, or spray coating.[117,231,232] In contrast to the etched surface demonstrated by Liu et al., omniphobic meshes make use of the periodic microscale texture from the mesh along with nanoscale texture attained from surface functionalization with other textural elements such as nanoparticles and microparticles to form a hierarchically textured surface with randomly positioned reentrant elements. As an example, we have spray coated a stainless-steel mesh surface with ZnO tetrapods that have tapered arms ending in tips with nanoscale dimensions (Figure C).[117] These randomly oriented reentrant structures were then functionalized with 1H,1H,2H,2H-perfluorooctanephosphonic acid to form a plastronic omniphobic surface that was manufacturable at scale and able to glide heavy crude oil. This robust, hierarchically textured surface produced contact angles of 160 ± 1° for water and 156 ± 1° for heavy crude oil.[117]

To produce reentrant, hierarchically textured surfaces with excellent adhesion, we have devised an alternative strategy directed at stabilizing a porous surface embedding a high density of surface plastrons. In particular, we have used electroless nickel plating to codeposit a nickel phosphide thin film in conjunction with poly(tetrafluoroethylene) beads (Figure B).[209] The evolution of hydrogen during electroless plating gives rise to a surface pitting, thereby creating an abundance of textural elements under the appropriate process conditions. Upon functionalization with 1H,1H,2H,2H-perfluorooctanephosphonic acid, such a combination of nanoscale texture and reentrant plastronic geometries yielded an omniphobic surface that was capable of facilitating drag reduction of up to 17% for a light hydrocarbon, castor oil, in laminar flow.[209] However, plastrons were observed not to survive high flow pressures or turbulent flow conditions, underscoring the need to develop more robust crosslinked architectures. An alternative electrochemical approach involves anodization of an aluminum surface to develop surface texture followed by the self-assembly of 1H,1H,2H,2H-perfluorodecyl-trichlorosilane, which yields a surface with a rapeseed oil contact angle of 150°.[233]

Another strategy that we have recently demonstrated uses colloidal crystal templating to define a doubly reentrant surface using polystyrene microspheres as sacrificial templates. An inverse opal geometry is constructed from sintered TiO2 nanospheres and upon functionalization with 1H,1H,2H,2H-perfluorooctanephosphonic acid yields a hierarchically textured, reentrant structure that glides heavy crude oil (achieving contact angles as high as 161 ± 2°) (Figure D).[218]

A superoleophobic coating can be created using a variety of techniques, as long as the aforementioned key components of reduced surface energy, hierarchical texturation, and doubly reentrant curvature are maintained. Approaches for casting polymeric films such as electrospinning, templating, layer-by-layer deposition, electrochemical deposition, plasma treatment, and self-assembly provide a means of inducing surface texture and have additional advantages such as flexibility, solution-phase processability, tunability of domain geometries based on molecular constituents, and applicability at scale.[234,226,227] Kumar et al. engineered hierarchical textured and nonwoven porous fibers using electrospun poly(1,6-heptadiyne) fibers modified with acrylonitrile butadiene styrene (ABS) that demonstrated hydrophobic behavior, robust thermal properties, and low ice adhesion.[235] Mülazim et al. have similarly explored the deposition of a photocured fluoroacrylate resin (Fluowet AC812, CH2=CHCOOC2H4(CF2CF2)n, where n = 3–6) with trimethylolpropane tris(3-mercaptopropionate), triallyl 1,3,5-triazine-2,4,6(1H,3H,5H)-trione, and silica particles treated with hexamethydisilazane, on ABS and high-impact polystyrene (HIPS).[236] These researchers observed a direct correlation between increasing hydrophobicity and oleophobicity and higher fluorine content.

Plasma nanotexturing represents an additional method to engineer oleophobic coatings. Ellinas et al. have demonstrated a superoleophobic/superhydrophobic PMMA surface prepared from the assembly of PS colloidal microparticles, subsequent plasma etching, and deposition of a fluorocarbon thin film.[237] The design of engineered surfaces that exhibit a high degree of oil repellency, excellent adhesion strength, resilience to corrosive species, and robustness to thermal cycling based on interrogation of physical mechanisms and elucidation of chemical design principles will bring disruptive innovation to the transportation and storage of heavy crude oil and bitumen. While geometric and chemical descriptors have emerged from the studies of lithographically patterned substrates and understanding of the interfacial structure, future research is focused on manufacturability of textured coatings and improving their resilience toward abrasive and corrosive species. Application at scale will further require the development of facile surface preparation procedures to enable retrofitting applications on existing infrastructure, or alternatively, the design of liners and sleeves that can be fitted onto the components.

Solid-Phase Transportation of Bitumen

To mitigate energy-intensive and wasteful processes of dilution and heating for transportation via pipelines,[152,238,239] as well as to mitigate the catastrophic environmental impact of oil spills on vulnerable ecosystems, much recent attention has focused on partial solidification of viscoelastic bitumen. Solid prills of bitumen can be much more safely transported using railroads, freight carriers, or container ships. A notable innovation from the Canadian National Railway Company (CN) comprises mixing and encasing bitumen within hydrocarbonaceous polymer shells.[240] The semisolid pucks are water-resistant, nonadhering, and buoyant; as such, this CanaPux technology has the potential to considerably reduce the cost and environmental risks associated with transporting liquid bitumen. However, the polymer shell wrapping the bitumen requires further processing at the point of use, and considering the massive volumes that need to be transported on a daily basis, inevitably necessitates the use of a large volume of polymeric media. An alternative method developed by BitCrude and Solideum BITTS involves the separation of the lighter fraction of bitumen and polymerization of viscous fractions under temperature or pressure to create bitumen bricks and balls.[22,241,242] The lighter fractions and bitumen balls are then transported separately as liquid- and solid-phase materials, respectively. This approach does not require the incorporation of extrinsic coating materials.

Recently, in collaboration with Cenovus Energy, we have developed a safe and viable method to form solid-phase bitumen microcapsules without the need for any extraneous additives. In this approach, the bitumen is reconfigured such as to encapsulate lighter fractions (saturates, aromatics, resins) of bitumen within a shell constituted from crosslinked asphaltenes. Asphaltenes, conjugated π-systems with aliphatic tails and various pendant functional groups (representative structural models of asphaltenes are shown in Figure A), can be readily extracted from bitumen by precipitating with n-heptane or hexanes. The asphaltenes are coated to form a shell enrobing deasphalted bitumen droplets in a core–shell configuration. Two different approaches have been developed to achieve the microencapsulation of bitumen. In the first approach, an automated jetting system equipped with a concentric nozzle facilitates the preparation of solid-phase bitumen microcapsules. The inner flow stream of concentric nozzles comprises heated bitumen, whereas the outer flow stream comprises a dispersion of asphaltenes in chlorinated organic solvents, as shown in Figure B. The flow stream is mechanically disrupted with an attenuator to enable the formation of uniform microcapsules. The microcapsules are collected in a water bath containing 2 wt % surfactant (poly(ethylene oxide)–poly(propylene oxide)–poly(ethylene oxide)) to prevent coalescence.[23] In the second approach, powdered ground asphaltenes are conformally coated onto hot bitumen droplets jetted from a singular flow nozzle onto a granular powder bath, as shown in Figure B. This enables thermally mediated crosslinking of asphaltenes around hot bitumen droplets.[24] In both cases, the high propensity of asphaltenes to aggregate, which represents a major source of flow impediments in reservoirs, is harnessed to facilitate crosslinking into a solid form that can be utilized for midstream transportation. The crosslinking of asphaltenes within the shells of microcapsules is a result of π–π interactions, metal–ligand interactions, thiol-ene reactions, Diels–Alder reactions, and disulfide formation.

Figure 12

(A) Representative structural models for asphaltenes, i.e., archipelago and island model. (B) Schematic depiction of the microencapsulation process of bitumen within asphaltene shells using a concentric and single nozzle. (C) Ensemble of laboratory-prepared microcapsules. (D) Confocal image of (i) a microcapsule from the top and (ii–iv) cross-sectional view of microcapsules. (E) Rupture of a microcapsule under mechanical stress. Images are reproduced with permission from Reprinted (adapted) with permission from Anita; Zaheer, W.; Douglas, L.; Sellers, D. G.; Gupta, S.; Banerjee, S., Asphaltene Microencapsulation of Bitumen as a Means of Solid-Phase Transport. Energy & Fuels 2021, 35 (8), 6576–6584. Copyright 2021 American Chemical Society. Reprinted from Fuel, 302, Anita; Zaheer, W.; Jakhar, K.; Antao, D. S.; Gupta, S.; Banerjee, S., Powder bed coating of bitumen with asphaltenes to obtain solid prills for midstream transportation, 121093, 2021, with permission from Elsevier.[23,24]

The recovered microcapsules have been examined by confocal microscopy to differentiate between the bitumen core and asphaltene shell. The darker regions in Figure D correspond to solidified shells comprising solidified asphaltenes, whereas fluorescent regions (Figure D(ii–iv)) in the cross-sectional image show lighter bitumen fractions that have a pronounced luminescent response. Since no extraneous chemical additives are present in solid-phase bitumen microcapsules, they can be readily fluidized into the original mixture through the application of compressive stress at a refinery, as shown in Figure E.

Solidification approaches that transform liquid bitumen into microcapsules, pucks, pellets, or bricks for transportation greatly expand opportunities for midstream transportation beyond pipelines and tanker cars to trucks, conventional freight cars, and container ships. As such, these represent a promising alternative to pipeline capacity constraints, energy-intensive separation processes inevitable upon dilution, and mitigating environmental hazards associated with oil spills.

Conclusions and Future Outlook

In this perspective article, several aspects of materials science challenges in the utilization of heavy crude oil have been highlighted with an emphasis on the Canadian Oil Sands. In particular, we have delineated the foundational principles underpinning candidate solutions for thermal insulation, membrane-assisted separations, corrosion protection, and midstream bitumen transportation. Our purpose is not to provide an exhaustive account of all of the challenges encountered in the extraction, transportation, and processing of heavy crude oil but to frame the need for surface modification and materials’ design approaches with a view toward the preservation of function under extreme environments commonly observed during operations. The demanding conditions for materials and surfaces are directly traceable to the high viscosity, low surface tension, and substantial sulfur content of heavy crude oil, which necessitates extensive thermal processes, warrants dilution/emulsification to ease flow, and engenders the need to protect corrodible components. The US Energy Information Administration’s predictions until 2050 demonstrate extensively and perhaps increasing worldwide reliance on heavy oil and unconventional deposits in general.[1] The materials’ technologies discussed in this perspective are of paramount importance to the economic viability of heavy oil production within the Canadian Oil Sands.

Thermal insulation is pivotal to overall energy efficiency, underpins protection of the infrastructure from degradation from thermal fatigue and is essential to maintain flow streams across the SAGD, emulsion separation, and midstream transportation processes. Tailored membrane technologies provide a means of decreasing reliance on energy-intensive thermal methods and hold promise for the separation of hydrocarbons from complex emulsions, engendering the separation of bitumen–diluent mixtures, and even fractionation of the different fractions of bitumen. As such, membrane technologies could usher in a new era of distributed chemical manufacturing and reduce reliance on large-scale refineries. Since the oil and gas infrastructure is predominantly constructed from low alloy steels, corrosion inhibition remains an existential challenge that is central to maintaining the continuity of operations and preventing catastrophic failures that can discharge hydrocarbons into the environment. Incorporating multiple modes of corrosion inhibition in protective coatings along with real-time monitoring of the integrity of components using advanced sensor technologies and machine learning methods holds promise for increased safety of the midstream transportation infrastructure. Limitations on accessible pipeline capacity have led to a massive surge in interest in surface modification and encapsulation technologies, spanning the range from drag reduction of oil flow in pipelines, oleophobic surfaces that readily glide viscous oil with no detectable residues, and encapsulation technologies that enable solid-phase transport. Such technologies will greatly reduce reliance on light condensate diluents, alleviate surface fouling and associated needs for cleaning, reduce waste, and cut down on energy expenditures.

Looking to the future, three strands of research hold particular promise. As the field of petroleomics attains maturity and detailed information of speciation and molecular structure of heavy oils becomes available, the rational design of approaches to engender separations, preclude wettability, and drive selective reaction chemistry during extraction and immediate postprocessing is becoming increasingly viable.[243−245] Such an idea of “refinery at the hole” combined with a push toward distributed chemical manufacturing can enable a new richly intertwined paradigm for the interconnected oil, plastics, and chemical industries. A particular frontier involves elucidation of interfacial structure at metal, mineral, and aqueous interfaces through spectroscopic methods, X-ray/neutron scattering, and first-principles calculations. Elucidation of interfacial structure will inform the design of methods to perturb and modify interactions, thereby enhancing oil recovery, aiding in fractionation of oleic mixtures, and informing the design of flow systems. Reactive flow systems that implement partial upgradation (such as desulfurization, hydrogenation, etc.) hold great promise for improved operations in midstream transportation and for obtaining better values at the refinery.

A second major area with the potential to transform midstream transportation and processing has its origins in decreasing levelized costs of electricity, which render electrochemical processes readily accessible on site for purposes such as hydrogen generation from electrocatalysis of water, electrochemical separations, electropolymerization, and direct electrochemical transformation of heavy oil fractions into value-added products. The ability to source hydrogen and power in remote locations as a result of advances in solar photovoltaics and photocatalysis has the potential to shift a substantial amount of the processing to the sites where heavy oil are extracted, greatly mitigating the challenges with midstream transportation, potentially bypassing the use of refineries for some fractions, and overall enabling access to a broader cross-section of refineries by achieving partial upgradation and separation.

A third major strand of research that holds the promise for accelerated design of materials for function in extreme environments involves the confluence of data science, machine learning, and artificial intelligence.[246−249] These approaches are particularly useful for the simultaneous optimization of multiple objectives, e.g., superoleophobicity and adhesion strength, along a Pareto frontier. Statistical learning approaches have the ability to enable the more systematic and strategic navigation of high-dimensional design spaces. In this context, such approaches can quantify the extent to which different parameters elicit a specific response (and identify null variables), rank features in order of their importance, and establish correlations across the different variables. The oil and gas industry has large data sets such as diagnostics information from infrastructure, extensive long-term testing data, and multimodal analysis of fluids and scale products. These data sets are particularly amenable to the use of statistical learning methods, which can analyze trends, estimate the magnitude of influence, and map correlations across diverse variables. In the context of materials’ design for the applications in heavy oil considered in this perspective, machine learning and sequential learning algorithms enable the construction of a response surface and its iterative improvement based on the identification of the next most valuable experiments. As such, these methods hold promise for accelerating convergence on the desired set of surface chemistries and microstructures by counterbalancing trade-offs between samples that would help to improve the extent to which the model captures the data set (exploration mode) or converge toward a predicted maximum or minimum within the response surface (exploitation mode). By mapping trends within the design space, statistical learning methods can thus help determine to identify specific metrics to optimize in subsequent experiments. Such sequential learning methods hold promise to entirely transform materials’ selection, prototyping, and piloting, enabling much more rapid deployment of new materials in the oil and gas infrastructure.